The transformational-generative concept of language learning as constructing a theory about language inspired this experiment. Language performance developed by a method of language learning designed to help the student study about a second language through manipulative exercises was compared to language performance developed by a method not directly designed to encourage theorizing about language. In this experiment the second language was English and the particular area of language capability tested was nominalization, using Vendler's typology, as a measure of maturity of language performance. Statistics derived from written tests by 10 students in the experimental method and 10 students in a control group, measured against the test results of 10 volunteer native speakers, indicated that the experimental study of nominalization through sentence-combining exercises did facilitate the development of language competence in that area and that it increased the capability of the experimental group toward the performance of the native speaker. Tables demonstrating the statistical results and an appendix on Vendler's typology of nominals are included. (VM)